A one-dimensional moving grid solution for the coupled non-linear equations governing multiphase flow in porous media. 1: Model development

A straightforward moving grid finite element method is developed to solve the one-dimensional coupled system of non-linear partial differential equations (PDEs) governing two- and three-phase flow in porous media. The method combines features from a number of self-adaptive grid techniques. These techniques are the equidistribution, the moving grid finite element and the local grid refinement/coarsening methods. Two equidistribution criteria, based on solution gradient and curvature, are employed and nodal distributions are computed iterativcly. Using the developed approach, an intermingle-free nodal distribution is guaranteed. The method involves examination of a single representative gradient to facilitate the application of moving grid algorithms to solve a non-linear coupled set of PDEs and includes a feature to limit mass balance error during nodal redistribution. The finite element part of the developed algorithm is verified against an existing finite difference model. A numerical simulation example involving a single-front two-phase flow problem is presented to illustrate model performance. Additional simulation examples are given in Part 2 of this paper. These examples include single and double moving fronts in two- and three-phase flow systems incorporating source/sink terms. Simulation sensitivity to the moving grid parameters is also explored in Part 2.     

Automatic grid generation of complex geometries in Cartesian co-ordinates

A method of automatic grid generation for complex boundaries in Cartesian co-ordinates is proposed in this paper. In addition to the Cartesian grid lines the diagonal segments are used for the approximations of complex geometries in Cartesian co-ordinates. A structured Cartesian grid is employed for the sake of the numerical simplicity and the potential of automatic grid generation. The automatic grid generation is achieved by this diagonal Cartesian method and the accuracy estimations of geometry approximations are given. The approximations of a few complex geometries, such as the multibody system in porous media, lake banks, grooved channels and spheres are shown and analyzed. The proposed method is verified by the numerical solutions of a rotated cavity flow. It is shown that the diagonal Cartesian method improves both the accuracy of geometry approximations and the numerical solution of a rotated cavity flow, comparing with the traditional saw-tooth method in which only Cartesian grid lines are utilized for geometry approximations. The stability and convergence of the proposed method is demonstrated. Finally, the application of the diagonal Cartesian method for the prediction of a grooved channel flow is presented. © 1998 John Wiley & Sons, Ltd.     

Control functions and grid qualities measurements in the elliptic grid generation around arbitrary surfaces

This paper describes the three‐dimensional elliptic grid generation. The two‐dimensional approach for the control functions obtained by modifying the Thomas–Middlecoff method is applied on the planes perpendicular to the main flow direction co‐ordinate, which is assumed to be the function of only one corresponding co‐ordinate in the computational domain. The grid orthogonality is improved by about 40 per cent compared with that of the algebraic initial grid. Copyright © 2000 John Wiley & Sons, Ltd.     

采用Poisson方程生成曲线网格时源项P、Q选择的研究

采用 Poisson方程进行曲线网格生成时 ,如何确定合适的调节因子 P、Q函数是网格生成技术中的一个重要的研究内容。本文提出一种新的构造 P、Q函数的方法 ,该方法直接利用边界网格节点的分布信息来控制区域内部网格节点的分布 ,生成的正交曲线网格令人满意。该方法可应用于河道、湖泊等一类复杂边界的二维流速场的数值模拟中。     

直接控制网格节点分布的曲线网格生成技术研究

指出了Thompson与Thomas曲线网格生成方法中控制网格分布的调节函数的问题所在,克服了Thomas曲线网格生成法中边界处局部线性化近似假定的缺陷,经过严格推导得出一组新的调节函数P、Q的表达式,并给出了曲线网格生成实例.实例检验表明,该调节函数能够对复杂边界的单连通域或多连通域生成理想的曲线网格,即边界处网格正交,内部网格分布能够适应物理量场变化的情形.在实际水利工程流场数值模拟中,该方法能够准确地使用边界条件,提高求解的精确度.     

A one-dimensional moving grid solution for the coupled non-linear equations governing multiphase flow in porous media. 2: Example simulations and sensitivity analysis

This paper presents numerical examples for the moving grid finite element algorithm derived in Part Ito solve the non-linear coupled set of PDEs governing immiscible multiphase flow in porous media in one dimension. Examples include single- and double-front simulations for two- and three-phase flow regimes and incorporating a mass sink. The modelling approach is shown to achieve significant savings in computation time and memory allocation when compared with fixed grid solutions of equivalent accuracy. This work includes sensitivity analyses for the parameters which are incorporated in the grid adaptation method, including the curvature weights, artificial viscosity and artificial repulsive force. It is found that the curvature weights are exponential functions of the negative ratio of the square root of the domain length to the number of discrete nodes. These weighting parameters are also shown to depend upon the shape of the front. On the basis of the examined simulations, it is recommended that artificial viscosity be neglected in the solution of the coupled non-linear set of PDEs governing multiphase flow in porous media. Similarly, use of a repulsive force is found to be unnecessary in simulations involving the migration of two liquid phases. For multiphase flows incorporating a gas phase it is recommended to use a non-zero value for the repulslive force to avoid development of an ill-conditioned nodal distribution matrix. An equation to evaluate the repulsive force under these circumstances is suggested.     

A modified adaptive grid method for recirculating flows

In this study a method of equidistribution of a weight function for grid adaption is modified to produce a smoother grid which yields a more accurate solution. In the original scheme the weight function was estimated on each grid independently and a large variation in the values of the, weight function could generate a highly skewed and non-uniform grid which produced large errors. In this study the weight function is smoothed by coupling neighbouring weight functions. Abrupt changes in the weight function are alleviated and a smoother grid distribution is obtained. With relatively minor modifications of the original weight function it is demonstrated in this study that the solution can be improved. The test cases presented are the one-dimensional convection-diffusion equation, a laminar polar cavity flow, a laminar backwardfacing step flow and a turbulent reacting sudden expansion pipe flow. Numerical efficiencies ranging from factors of five to 10 are achieved over uniform grid methods.     

Algebraic–hyperbolic grid generation with precise control of intersection of angles

A new approach to construct high quality meshes for flow calculations is investigated in the context of algebraic grid generation, where the angle of intersection between two alternate co‐ordinate lines can be specified by the user. From an initial grid, new co‐ordinate lines in one parameter direction are determined by solving a set of ordinary differential equations (ODE). These trajectories intersect the alternate co‐ordinate lines of the initial grid in a prescribed angle. The resulting grid function is proven to be folding‐free. The advantages with respect to grid quality achieved by prescribing the angles are gained at the expense that the boundary distribution can only be prescribed on three of four boundaries. In view of a sparse representation of the grid function, the points of intersection are interpolated by a B‐spline surface. Thus, the grid can be accessed evaluating the generated parametric representation of the computational domain. Investigations of several test configurations have shown that useful grids with high resolution can be computed from the B‐spline representation only depending on a small number of locally chosen parameters. Therefore, only a small amount of computational memory is required for storing these parameters and evaluating the grid function can be efficiently performed at reduced computational costs. This is particularly promising when the mesh has to be frequently changed or updated in time. Copyright © 2000 John Wiley & Sons, Ltd.     

Boundary conformed co-ordinate systems for selected two-dimensional fluid flow problems. Part I: Generation of BFGs

In this paper the generation of general curvilinear co-ordinate systems for use in selected two-dimensional fluid flow problems is presented. The curvilinear co-ordinate systems are obtained from the numerical solution of a system of Poisson equations. The computational grids obtained by this technique allow for curved grid lines such that the boundary of the solution domain coincides with a grid line. Hence, these meshes are called boundary fitted grids (BFG). The physical solution area is mapped onto a set of connected rectangles in the transformed (computational) plane which form a composite mesh. All numerical calculations are performed in the transformed plane. Since the computational domain is a rectangle and a uniform grid with mesh spacings Δξ = Δη = 1 (in two-dimensions) is used, the computer programming is substantially facilitated. By means of control functions, which form the r.h.s. of the Poisson equations, the clustering of grid lines or grid points is governed. This allows a very fine resolution at certain specified locations and includes adaptive grid generation. The first two sections outline the general features of BFGs, and in section 3 the general transformation rules along with the necessary concepts of differential geometry are given. In section 4 the transformed grid generation equations are derived and control functions are specified. Expressions for grid adaptation arc also presented. Section 5 briefly discusses the numerical solution of the transformed grid generation equations using sucessive overrelaxation and shows a sample calculation where the FAS (full approximation scheme) multigrid technique was employed. In the companion paper (Part II), the application of the BFG method to selected fluid flow problems is addressed.     

Elliptic adaptive grid generation and area equidistribution

The derivation of elliptic adaptive grid control functions which satisfy the area equidistribution concept is presented. The resulting expressions are derived without approximation and are shown to provide explicit control over cell area distributions. A modification to the equidistribution concept which yields control functions that enable additional control of the near‐boundary grid resolution is also proposed. A computer code which incorporates these control functions has been developed and is applied to a series of complex fluid flows to demonstrate the validity and utility of the derived expressions. Copyright © 1999 John Wiley & Sons, Ltd.     

多尺度嵌入式离散裂缝模型模拟方法

天然裂缝性油藏和人工压裂油藏内裂缝形态多样,分布复杂,传统的离散裂缝模型将裂缝作为基岩网格的边界,采用非结构化网格进行网格划分,其划分过程复杂,计算量大。嵌入式离散裂缝模型划分网格时不需要考虑油藏内的裂缝形态,只需对基岩系统进行简单的网格剖分,可以大大降低网格划分的复杂度,从而提高计算效率。然而,在油藏级别的数值模拟和人工压裂裂缝下的产能分析中,仍然存在计算量巨大、模拟时间过长的问题。本文提出嵌入式离散裂缝模型的多尺度数值计算格式,使用多尺度模拟有限差分法研究嵌入式离散裂缝模型渗流问题。通过在粗网格上求解局部流动问题计算多尺度基函数,多尺度基函数可以捕捉裂缝与基岩间的相互关系,反映单元内的非均质性,因此该方法既有传统尺度升级法的计算效率,又可以保证计算精度,数值结果表明这是一种有效的裂缝性油藏数值模拟方法。     

The grid method

Determining strain with grids is one of the oldest and simplest methods of experimental stress analysis. Here, the method is reviewed. Various techniques that have been developed to print, record and analyze grids are discussed, and the types of problems to which the grid method has been applied are presented.     

A finite volume method for numerical grid generation

A novel method to generate body‐fitted grids based on the direct solution for three scalar functions is derived. The solution for scalar variables ξ, η and ζ is obtained with a conventional finite volume method based on a physical space formulation. The grid is adapted or re‐zoned to eliminate the residual error between the current solution and the desired solution, by means of an implicit grid‐correction procedure. The scalar variables are re‐mapped and the process is reiterated until convergence is obtained. Calculations are performed for a variety of problems by assuming combined Dirichlet–Neumann and pure Dirichlet boundary conditions involving the use of transcendental control functions, as well as functions designed to effect grid control automatically on the basis of boundary values. The use of dimensional analysis to build stable exponential functions and other control functions is demonstrated. Automatic procedures are implemented: one based on a finite difference approximation to the Cristoffel terms assuming local‐boundary orthogonality, and another designed to procure boundary orthogonality. The performance of the new scheme is shown to be comparable with that of conventional inverse methods when calculations are performed on benchmark problems through the application of point‐by‐point and whole‐field solution schemes. Advantages and disadvantages of the present method are critically appraised. Copyright © 1999 National Research Council of Canada.     

A simple grid generation method

This paper presents a simple grid generation method which adopts the uni-directional interpolation idea but only interpolates one co-ordinate between two opposite boundaries. The use of this new scheme showed that (a) it is convenient to use; and (b) compared with the built-in grid generator of certain commercial computational fluid dynamics (CFD) code, it saves time significantly. © 1998 John Wiley & Sons, Ltd.     

一种新的准结构网格生成方法

在有限元分析中,高质量的结构网格可以有效地提高有限元分析的精度,但结构网格的几何适应性差,针对复杂边界的二维计算模型,现有的方法很难自动生成高质量的结构网格;而非结构网格几何适应性很好,但存在计算效率低和精度差等问题。提出了一种新的准结构网格生成方法,能够实现复杂区域的网格自动生成并且具有高网格质量。该方法首先对计算区域运用Delaunay三角剖分技术生成粗背景网格;然后利用背景网格,使用优化的Voronoi图生成过渡的蜂巢网格;最后,通过中心圆方法对蜂巢网格单元进行结构网格剖分。分析NACA0012翼型数值模拟结果表明,提出的新准结构网格生成方法能够对边界复杂的模型自动生成高质量的网格,并且通过三种不同拓扑类型网格计算结果相互对比及与实验结果对比,证明准结构网格具有高计算精度。     

非结构混合网格高超声速绕流与磁场干扰数值模拟

对均匀磁场干扰下的二维钝头体无粘高超声速流场进行了基于非结构混合网格的数值模拟.受磁流体力学方程组高度非线性的影响及考虑到数值模拟格式的精度,目前在此类流场的数值模拟中大多使用结构网格及有限差分方法,因而在三维复杂外形及复杂流场方面的研究受到限制.本文主要探索使用非结构网格(含混合网格)技术时的数值模拟方法.控制方程为耦合了Maxwell方程及无粘流体力学方程的磁流体力学方程组,数值离散格式采用Jameson有限体积格心格式,5步Runge-Kutta显式时间推进.计算模型为二维钝头体,初始磁场均匀分布.对不同磁感应强度影响下的高超声速流场进行了数值模拟,并与有限的资料进行了对比,得到了较符合的结果.     

Simple treatment of non‐aligned boundaries in a Cartesian grid shallow flow model

A simple method is proposed for treating curved or irregular boundaries in Cartesian grid shallow flow models. It directly evaluates fictional values in ‘ghost’ cells adjacent to boundary cells and requires no interpolation or generation of cut cells. The boundary treatment is implemented in a dynamically adaptive quadtree grid‐based solver of the hyperbolic shallow water equations and validated against several test cases with analytical or alternative numerical solutions. The method is easy to code, accurate, and demonstrably effective in dealing with irregular computational domains in shallow flow simulations. Results are presented for still water in a basin of complicated geometry, steady hydraulic jump in an open channel with a converging sidewall, wind‐induced circulation in a circular shallow lake, and shock wave diffraction in a channel containing a contraction and expansion. Copyright © 2007 John Wiley & Sons, Ltd.     

A self-adaptive gridding for inviscid transonic projectile aerodynamics computation

An adaptive grid generation technique based on modified variational principles coupled with an exponential clustering has been developed and tested successfully for the computation of steady inviscid transonic projectile aerodynamics. The isoperimetric problem for adaptive gridding is to extremize a grid smoothness functional subject to grid orthogonality and resolution functionals; however, the Lagrange multipliers have been assumed to be variables with zero variation and are properly chosen as functions of local grid size to enhance locally the grid resolution as well as to maintain the weight of three grid characteristics the same over the entire flow field. With computed pressure gradient as the control function for grid adaptation, the resulting Euler equations cannot provide sufficient grid resolution in the boundary layer region of the projectile geometry; hence, a clustering technique is needed to redistribute the points along the normal grid lines. A grid generation code has been developed and coupled to an axisymmetric thin-layer Navier-Stokes code for self-adaptive grid generation. For the three transonic flow cases considered, M_∞ = 0.91, 0.96 and 1.10, the distribution of surface pressure calculated from the inviscid option of the Navier-Stokes code is indeed in excellent agreement with published measured data.     

A numerical procedure for the generation of orthogonal body-fitted co-ordinate systems with direct determination of grid points on the boundary

The procedure proposed is based on the solution by finite difference means of a set of Laplace's equations, by the application of a relaxation method. The curvilinear orthogonal grid so generated is fitted to a 2-D physical domain with closed boundary and the contribution of the present work consists in the arbitrary choice of grid points on two adjacent boundaries, in order to achieve the desired density of grid points where the geometry of the boundaries varies rapidly. The method proposed is rapid and stable. Some characteristic examples are finally presented.     

An algebraic grid generation technique for time-varying two-dimensional spatial domains

An efficient and versatile algebraic grid generation technique is presented for generating grid points in irregularly shaped and time-varying spatial domains. The method presented is based on the ‘two-boundary technique’ of Smith. The usefulness and the feasibility of the grid generation technique were demonstrated by (1) generating grid points inside one of the combustion chambers of a motored two-dimensional rotary engine and (2) obtaining numerical solutions for the flow field inside one of those combustion chambers.